Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 84540
    Prev
    Next
    INDEX
    Explanations

    separate

    np_max-act · gemini-2.0-flash

    The neuron activates specifically on occurrences of the word "separate."

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    リン
    -0.06
    field
    -0.06
     Think
    -0.06
    -function
    -0.06
    ्रमण
    -0.06
    こんに
    -0.06
     His
    -0.06
     Robertson
    -0.06
     ś
    -0.06
    ソン
    -0.06
    POSITIVE LOGITS
     separate
    0.14
     Separate
    0.11
     separately
    0.10
     own
    0.08
     special
    0.08
     Gather
    0.08
    另
    0.08
     seperate
    0.08
    RGBO
    0.08
    emsp
    0.07
    Activations Density 0.008%

    No Known Activations